Dell Technologies World 2025 | Alex Barretto & Amit Sawney, Dell Technologies

Clips
More from Dell Technologies World 2025

Alex Barretto

SVP

Dell Technologies

Amit Sawhney

VP

Dell Technologies

play_circle_outline Maximizing AI Integration in Hardware and Software: Enhancing Go-To-Market Strategies and Supply Chain Services

play_circle_outline Data mesh unification and proprietary data sets

play_circle_outline Digitizing core processes for AI amplification

play_circle_outline Dell Tech World coverage

Info
Transcript

Alex Barretto & Amit Sawney, Dell Technologies

Alex Barretto

SVP Dell Technologies

Amit Sawhney

VP Dell Technologies

Alex Barretto, senior vice president at Dell Technologies Inc., and Amit Sawhney, vice president of technology and portfolio at Dell Technologies Inc., join theCUBE’s Savannah Peterson and Dave Vellante at Dell Technologies World 2025 to discuss how Dell is embedding AI across its portfolio. Their conversation explores how rapid experimentation and data governance are fueling smarter, more agile innovations.

Barretto and Sawhney detail how AI is reshaping both hardware and software development across Dell’s product ecosystem. From tailored customer e... Read more

explore Keep Exploring

What areas are being focused on behind the scenes to improve products and services, and how is AI being used to drive differentiated outcomes for customers? add

What is the AI strategy of the company and how has it evolved to 2025? add

What strategy did Alex and I talk about internally that involved digitizing core processes, simplifying and standardizing them, automating them, and then implementing AI models on top of them for amplification and speed? add

What is the advantage of running and deploying models quickly and easily using AI PCs in the field of technology? add

bolt Powered by CUBE AI

Alex Barretto & Amit Sawney, Dell Technologies

search

>> Good afternoon, Dell fans, and welcome back to fabulous Las Vegas, Nevada. We're here midway through day three of our coverage of Dell Tech World. My name's Savannah Peterson, here with Dave Vellante in the cockpit, learning from all the smartest people all week. We're killing it. We've got a lot of hardware we're showing off today.

>> I know. Have we talked about AI yet?

>> What does that mean? What does that acronym stand for, Dave? I'm not exactly sure. You know who I think could tell us are Alex and Amit. Welcome to the show, guys, thank you so much for being here.

>> Thank you for having us.

>> Thank you for having us.

>> This is very exciting. So one of the things that gets me really excited about our coverage and analysis of Dell is you're truly an ecosystem player. It's not one slice of the pie. You're working with all the folks to do all the things to bring AI everywhere. So talk to me, Alex, when you came up, you said it's your team who builds these. Must be a very exciting time for you all to start getting them out in the wild.

>> Super, super exciting. I mean, you're trying a product yourself. It's amazing product, great battery life, really fantastic. But the best thing is how we're doing things behind the scenes, right? So we've got a lot of AI, as you said, Dave, going on and there are four areas that we're focused on. So go-to-market supply chain, services, and product. Two of those we're distinctively for my organization working on. So the way we envision products, the way we think about our hardware engineering, software engineering, injecting AI and how we actually do software engineering to make it better so we do it faster and better, putting solutions in front of our customers that are better solutions, actually delivering customer value. On the services side, same thing. We have a hundred million connected devices. Think about the power that telemetry coming back. Then we use that and harness AI to drive differentiated outcomes for our customers. It's a great tool that drives both productivity and at the same time customer value. How often do you see that? Right? Two, that actually does both at once.

>> Those things can often be competing, sometimes depending on the ecosystem. So it makes a lot of sense. You just described a lot of hard work right there. Four different areas that are very key. How do you balance prioritizing what you're working on and what you guys are going to produce next?

>> Yes, that's a great question. We actually are big fans of experimentation, so we're going across the four areas going very rapidly. We try lots of things. Not all of them work out immediately, but it's just try to experiment very fast and we learn very quickly. So it's really about speed. We're trying to do a lot of use cases very rapidly.

>> And that's never been more important than it is right now.

>> Exactly. And actually AI unlocks speed too, right? Because we can experiment with things rather rapidly and adapt very quickly. So the prioritization is really about speed.

>> It was interesting to hear Jeff's keynote yesterday. He said, we do ninety-day sprints, we have five primary use cases as Alex, you just said revenue, supply chain, go-to-market and services. And then he talked about this data mesh unification and all the proprietary data sets that you guys have in-house. So guys, a couple of years ago you announced this AI strategy, AI in, on, with, for. Maybe you could take us through that again and help us just bring it to 2025.

>> Yeah, I think a good example of that is thinking about what we're doing in services, what we call next best action, right? We're taking the power of our data.

>> Yeah, big theme of the week.

>> Exactly. And then using that data set in our products to really drive differentiated outcomes. So we are actually generating insights into the product itself so it can auto-heal and actually be very seamless experience from a usage standpoint. So this notion of in the past, having to call support when you have an issue and they go through an if-then-else statement and if this, then you do that. That's gone. That's gone. The product's taking care of the issues by itself. In the unlikely event, you actually have to reach out to somebody. That person actually has an AI agent working with them. It's a multi-variable equation drive a differentiated value. So it is great to see the power of that.

>> And it's a win-win scenario. If you think about it from a customer perspective, the speed is so much better and from Dell's standpoint, the value proposition is very high because we are able to serve the intelligence to the customer when they need it, how they need it, and it drives the value back into our ecosystem.

>> The other thing David, I think is super interesting is we started doing this internally from a productivity standpoint.

>> You all drink your own champagne folks, which I love.

>> That's right. Customer zero right here, that's right.

>> Which is so important.

>> I call it dog fooding. So don't hate me. Because it couldn't have been all pretty, it couldn't have been all verve.

>> Have you seen me after too much champagne? It's not pretty either, my friend.

>> The interesting thing is we actually have customers reaching out to us and say, "Hey, how did you do it? What did you learn? Because we actually want to learn from your experience." And it's amazing to see them, the number of use cases. So a simple service use case slightly tweaked actually applies to a variety of, so airlines, think of airlines in the airports and all the gate agents.

>> You mean my home?

>> There you go. So they could be actually much more efficient by using the power of AI. So it's actually unlocked a number of conversations with our customers and how they can modernize and they can adapt AI in driving their business.

>> Yeah. So tell me some of the learnings that you've had internally from this experimentation that have allowed you to accelerate what you're able to offer for your customers.

>> The key word already said, speed. Speed is everything. The 90-day spreads, like you just said, Dave, like Jeff said yesterday, those are important words because speed is everything. Rapid experimentation, learning from it, understanding what works, just positioning it, moving onto the next things, leveraging our infrastructure to the best of its potential, ensuring that the data, which is our absolute gold mine, our IP stays our control point. We continue to build upon it. So those are some of the great learnings that we have had with these experimentation and we've been able to scale it very fast, leveraging our IP, leveraging our infrastructure, leveraging our solutions and our services, which is what we can offer to the customers because we to your point, have drunk our own champagne and have learned from our own solutions, which we are ready to help every enterprise customer out there with.

>> What approach did you use to avoid what I call paving the cow path from a process standpoint? In other words, taking intelligence and putting it on top of a frankly crappy process or maybe a process that's outdated. Jeff actually said it yesterday.

>> He did.

>> What he said, he said, "Basically if you got shitty data, you got shitty AI."

>> We know that. Garbage in, garbage out, right?

>> Right. I mean that's true.

>> So how did you avoid that, that problem?

>> It actually is a super important point that Jeff made and you're alluding to, right? We can't just go plug in AI and expect that all things are going to be taken care of. The first step we took was actually looking at our process and say, "This doesn't even make sense. We have to reimagine our process and reimagine in a way that actually fits what with the desired outcome we're trying to drive to." So the first step was that, looking at our process, re-engineering, re-imagining how we do things, that was step one. And that quite frankly, we changed a lot of things. We actually fundamentally said, "Yeah, we want to do things differently." Not a little tweak, fundamental differences, and then we apply AI on top of that. We actually have a whole data engineering team just looking at that. How does the data actually map through the process? How should it be changed so we can get the value?

>> Alex and I joke internally about a layered cake or a pyramid strategy that we have, which is exactly what we just talked about, is you start with digitizing and digitalizing your core processes. That's what we did. We ensured that we were simplifying and standardizing our processes. We were automating it, like said yesterday by Jeff. On top of it, we brought the AI models because AI amplifies on a right set of digitized processes and obviously follows data. So once we get the processes digitized, we got the right data governance built upon it, and then we brought AI. That's where the amplification happened. That's where the speed was achieved and that's where we got the outcomes that we got.

>> You just said a word that I don't think the industry talks about enough, and that's standardization. It makes such a difference in terms of the velocity conundrum. So how did you get everybody on the same page, I guess is my question? How did you get the organization ready?

>> Yeah, it does take this notion of data governance. We talk a lot about that. Bringing everybody together and say, "This is how we will do it." To be totally direct, it's, "You get on the boat or-"

>> Or get left behind, baby. Yes. We're going.

>> You got to get on the boat. There's strict governance in how we approach data in our processes for that reason because we want it standardized. So nobody going off on a tangent. The other piece is that focus, right back to we are trying use cases around go-to-market supply chain services and product. So if you want to have a different tangential point and it doesn't fit one of these fours, we're like, "That probably doesn't make sense." Because that's where we extract value, that's where our customers expect. So the governance is a very important aspect of it for sure.

>> It absolutely is. I'm curious, since this is so fun that your team, shout out to your team, we love our laptops. Since your team made these, what's the thing that you've done with your AI PC that gave you a wow moment or a light bulb moment?

>> Well, there's a lot of great things in these products, so thanks for the shout-out on that. I appreciate it. By the way, the one you're using, it's one of my favorites. So it's great to see and you both have it, so that's awesome. The way we actually think about the software engineering is actually very different. So traditionally, the way we approach everything from our firmware to BIOS to applications that we've built into the product is it used to be the more traditional approach, a little more of the waterfall even. We used to talk about being agile. We actually now are agile. We're moving at speed because of the power of AI, putting that quality code out very fast, that delivers the outcome to our customers. So when you use the product, you have a total experience, it's the hardware and the software together giving you that ultimate experience.

>> You know the first thing I did with my AI PC? I designed my nails for the show.

>> Excellent.

>> I opened up my laptop, got on co-pilot, uploaded the color palette for the show and all the stuff that I know about Dell.

>> That's a different use case I hadn't heard about.

>> That is an interesting use case. Another one, if you go walk 10 steps, you'll see a gigantic solutions being run on AI PCs right as we speak.

>> Yeah.

>> Yeah, they're right there.

>> It feels like this whole show floor is powered by your AI PCs, which is very fun. What about you, Amit? What's the most exciting thing for you about these?

>> For me, again, it's the power of running and deploying the models ease at speed. Because as technologists, we love to get those models out there quickly and AI PC is absolutely enabling that for us. Like I said, agentic solutions that we just built across product and services, they're running on AI PCs, they're running on NTU and GPUs, which if you think about it, the speed is immense computational speed. The way for us to learn from that, that compresses the cycle time, which allows us to drive to the outcomes faster. Back to the conversation about speed.

>> You think about the GB10, the GB300 that we announced two days ago. Those are, think about actually running now a large-lingual model on your device, literally sitting on your device. It unlocks immense value. So the products are fantastic.

>> So I'm thinking ahead a few years or maybe it's not going to be that fast. And so everybody's got an intelligent agents on their laptops. They've got them all inside the infrastructure. They're throughout the organization, AI-ified my organization. I imagine when you guys went through the process internally, and of course it's not over, it's still going through it, but there was initial, maybe it wasn't a super heavy lift, but there had to be a somewhat heavy lift, especially on the data side, bringing all the disparate data together, de-siloing, harmonizing that data. Okay, so let's assume in the future we've got our data states in much better shape. I'm envisioning a time when the agents, you got to do something. There's a task at hand or an initiative and you're trying to figure out the best process and the agents just-

>> Are doing it for you.

>> They're basically composable blocks that come together and actually create a new process and it's fast. Maybe there's some tweaks that a human in the loop wants to make.

>> I don't think this is a few years away. I think this is a couple months away.

>> So what has to happen for that to happen? How far is that away? Is it years? Is it months? Is it today?

>> I agree with you. So I don't think it's years away.

>> Yeah, I think we're real close.

>> I think we're very close to that. We are actually experimenting with that right now. As we said, we can look at some demos right now, but this notion of agents talking to each other to go do things, it's here today. And the next step is exactly what you said, Dave, them actually working together and create something new, right? That's the next evolution. But I think it's going to be very quickly, because you look at the amount of energy being spent on this and the development and the speed at which we are iterating through it. It's incredible.

>> And we are making it real. As we speak, we have these agents that are talking to each other with the A-to-A protocol. We have agents that are actually advising the other agents to take tasks. As you, again, you can see the product and the services agents are talking to each other. Those boundaries are blurred.

>> So I think about product returns. It's kind of an interesting use case. Why would you return a product? If I take go to... I go to Lowe's and I accidentally bought the wrong one or whatever.

>> It could be an extension cord you need to replace.

>> So I come back, could be an extension cord that had a little cut in it, whatever it was, and so inside joke. Okay, so sometimes it's pretty straightforward. Other times there might be an exception and you have to go to a human and sort of adjudicate, and so maybe this is available today, I'd be interested in your thoughts. Or maybe it's months away, maybe it's not. The agent learns from the reasoning traces of the human and then the next time that exception comes in, the agent takes care of it.

>> Exactly, yeah. I think that's well said. You have two important points there. One, it's not like it's a agent or human sphere. It is a hybrid, right? The workflows, the way they work is there are times that the human does get involved and obviously the agent's always there helping. So it's a hybrid sphere. And the second important point, it's a learning environment, so it's not like they're following a script and it's always that script. They're adapting based on the knowledge base, real time and getting better as we go. That is truly phenomenal, right? I mean, I remember the first time we used to talk about AI, it was truly just the if the now statement, now it's real learning. It's real change. It's amazing.

>> It's that Z-axis. There's a whole new world out there for everybody and what's going to be possible. Final question for you gentlemen. I'm actually very curious to hear what you have to say. When we're hanging out at Dell Tech world in 2026, what do you hope to be able to say then that you cannot yet say today?

>> Ooh, great question.

>> Alex. I'll start with you.

>> Sure. I actually would explore on this point, agentic. Really we have unleashed the value of agents where they're actually not only talking to each other, but they're generating net new ideas and processes. 2027 right here, we'll have that conversation.

>> Can't wait. Looking forward to it now. Looking forward to it. What about you, Amit?

>> I actually second that. I was going to say the same thing. It's like the value of agentic and the use cases that are being presented to us, we couldn't have imagined those few years back.

>> No.

>> Absolutely not.

>> No way.

>> So a year from now sitting here and telling you like XYZ can happen, that absolutely can happen. We will be shocked the value that agentic will bring.

>> Well, we look forward to telling those stories with you guys next year. Thank you so much for taking the time-

>> Thank you.

>> Thank you for having us....

>> this afternoon. Really appreciate it.

>> Thanks, you guys.

>> Thank you, Dave. Always fantastic insights. And thank all of you for tuning in wherever you might be. We're here in fabulous Las Vegas, Nevada at Dell Tech World. My name's Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.